Air Combat Autoresolve Formula Tweaks / Values

The autoresolve system has recently been made a bit more complex and should better reflect the realities of the air combat. However, there have also been some reports of the values being a bit off in the new system.

I thought I should explain exactly how the system works so people understand it, then throw the floor open to balance suggestions (balancing the existing values more than changing the calculation itself).

The Autoresolve essentially assigns the humans and aliens a score, then compares the two.

• The higher score wins the combat and destroys the enemy
  
• If the score is 50% higher than the losing score, the victors take no damage
  
• If less than 50% higher, the victors take proportional damage based on how much of that 50% they failed to achieve (40% higher means the victors take 20% damage)
  

The human / alien score is calculated as follows:

• The chassis of each fighter has an inherent strength value in aircrafts.xml
  
• The weapons on each fighter have a strength value set in aircraftweapons.xml
  
  • There is a different value to be used against targets that can Roll and those that cannot, which is used mostly for Torpedoes so they are not effective against fighters.
    
  • If a weapon is not at 100% ammo, its strength is modified by its ammo level.
    

  [*]Once all these are added up, a random roll of plus / minus 10% is added to the total for each side

  [*]The health of a combatant linearly modifies the score of chassis and weapons it carries.

  [*]If a combatant has less than 20% fuel, its score is modified linearly in proportion.

So, now that's all been explained, if there's instances where the autoresolve is giving incorrect / unrealistic results then I'm happy to hear proposed balance changes to the existing values for the weapons / chassis.

Which build is this implemented in?

[*]The health of a combatant linearly modifies the score of chassis and weapons it carries.

I'm not sure how I feel about this, as health doesn't have any impact on aircraft performance in game. So if I'm at 50% health, the fighter is 50% less effective in auto resolve? Which means it might lose auto resolve 100% of the time, but if I control it manually I could win 100% of the time. A better way might be to calculate the result regardless of health, and if the result kill's a fighter, recalculate excluding that fighter, or at that point, apply an effectiveness penalty to the fighter (so it still has an impact on the fight).

How does autoresolve calculate fuel and ammo usage? I've found in V21 Stable Steam that fuel usage is typically less, and ammo usage is typically more, than if I manually control.

Apologies for the repetition - and it's not 100% on topic - but with the calculation explicated its possible to highlight this issue with greater certainty:

A Foxtrot with Alenium Torpedoes should have a combat score of 900 (300 for the aircraft + 600 for the torpedoes). A Corvette has a combat strength of 1700. According to the formular outlined above, there's no way that the Foxtrot should have a 100% chance to win against a Corvette (which is what is displayed in the bug report related to this topic). Indeed, it should be 0% - even with randomisation the Corvette will always have a higher score than the Foxtrot, meaning it should always win. This implies then that there's an error with the calculation somewhere.

Auto-resolve doesn't consume fuel, but I think it does a rough approximation of the ammo required to kill the target. I forget exactly how it works.

But anyway, the above is correct as of V22 Experimental 5. However, I'm less interested in changing the mechanics than I am balancing what is already there.

@Kabill - the Corvette has two torpedoes. So it's 300 for the aircraft and 1200 for the torpedoes, giving 1500 and a max random roll of 1650.

This isn't high enough to give you victory on its own so something is going wrong, but it's pretty close. I wonder if the UFOs are suffering from the fuel modifier or something themselves?

@Kabill - the Corvette has two torpedoes. So it's 300 for the aircraft and 1200 for the torpedoes, giving 1500 and a max random roll of 1650.

This isn't high enough to give you victory on its own so something is going wrong, but it's pretty close. I wonder if the UFOs are suffering from the fuel modifier or something themselves?

Yes, of course. Sorry, for some reason I'd been treating the Foxtrot as having one hardpoint with two missiles in it rather than two hardpoints with one! Apologies for being dense.

Actually, the random roll for the Corvette could also be -10% so it's possible that the rolls might be working fine...although it shouldn't give a 100% victory chance in that case.

Do you get a 100% victory chance every time you attack a Corvette with a Foxtrot?

How is damage generated from autoresolve handled on aircraft with existing damage? Is it just tacked on?

------------------

Possible bug. So if I understand the calculation correctly, a MiG with an autoresolve value of 1200 verses a scout with a autoresolve value of 1000 should not only be 100% victory but also get away scot free. How does the double ko figure into this? Because that happens (it just happened to me!).

Auto-resolve doesn't consume fuel, but I think it does a rough approximation of the ammo required to kill the target. I forget exactly how it works.

Well, that makes sense why ti seems to use less fuel. Is there a plan to add fuel consumption? Seem like a bit of a cheat: if I need to chase down another UFO, and I have a high or 100% auto resolve chance, I can avoid fuel consumption from combat by auto resolving.

But anyway, the above is correct as of V22 Experimental 5. However, I'm less interested in changing the mechanics than I am balancing what is already there.

In that case, I'd just remove the health modifier, or reduce its impact.

I'm not sure how I feel about this, as health doesn't have any impact on aircraft performance in game.

But it does in a way if you consider that a UFO with half health only takes half the ammunition to kill so less risk to your own craft.

Think about a craft with 500 max strength but at 50% from damage for a total of 250.

A fighter with 300 strength would be unable to kill it (according to autoresolve) at full health but would be able to when it was damaged.

In the same way a single missile would not be able to kill a full health Corvette but it could finish off a damaged one.

The fight would be significantly easier as reflected in the auto resolve numbers.

@Chris - how does it work with multiple craft on each side?

For example attacking with Foxtrots against a Corvette and fighter escorts as the heavy missiles would have different values depending on which craft they were attacking.

How does the auto resolve decide which set of values should be used?

For me the balance seems fine but the basic mechanics could use a little tweaking.

For example if the potential total damage output of the human weapons doesn't at least equal the total health of the enemy ship then it should not be able to force a crash.

At best it should be able to make the enemy flee or break off after damaging the enemy to the max capacity of its weapons.

Fuel consumption could also be roughly calculated by using the fire rate of the weapons and the damage required to kill all enemy ships with a bit of leeway for travel time.

@Max - Remember the + / - 10% random rolls for each side. The scout can go up to 1100 and the MiG can go down to 1080.

@Muskrat - Health has plenty of effect on combat. Otherwise a UFO on 1% health would be considered just as much as a threat as one on full health. The relationship's not necessarily as linear as the formula uses, but there's a limit to how granular we can make the calculation.

@Gauddlike - If there's multiple craft, if any of them are not able to roll then the game assumes none of them are able to roll. The damage output of the weapons doesn't need its own calculation, it should be reflected in the scores of the weapons in the existing system.

@All - As I've said twice now, I'm not interested in changing the mechanics any further. We've got about a month until completion and we have bigger fish to fry. Fuel consumption doesn't matter because autoresolve still consumes ammo, which is the bigger limiting factor in most cases. Let's concentrate on the values (or at least fights that have unexpected outcomes), please.

So about that unexpected outcome I linked to which silencer posted up - the double ko where a MiG which has a larger autoresolve value than the Scout (1200 verses 1000), wins the combat but ends up with 0% health. That must be an unexpected resolution of the less-than-50% damage calculation.

But it does in a way if you consider that a UFO with half health only takes half the ammunition to kill so less risk to your own craft.

Think about a craft with 500 max strength but at 50% from damage for a total of 250.

A fighter with 300 strength would be unable to kill it (according to autoresolve) at full health but would be able to when it was damaged.

In the same way a single missile would not be able to kill a full health Corvette but it could finish off a damaged one.

The fight would be significantly easier as reflected in the auto resolve numbers.

@Muskrat - Health has plenty of effect on combat. Otherwise a UFO on 1% health would be considered just as much as a threat as one on full health. The relationship's not necessarily as linear as the formula uses, but there's a limit to how granular we can make the calculation.

I agree with what's being said here. My concern was from a my fighter being damaged stand point, as that's more likely to occur. Very rarely do I have the opportunity to re-engage already damages UFO's.

So does the health penalty only affect the fighter chasis score, irregardless of weapon score? For instance: Foxtrot with 2 alenium torpedoes (300 Aircraft + 1200 Weapon = 1500). At 50% health is it's score 1350 (300*0.5 + 1200), or is it 750 (1500*0.5)? If 1350, I think that's ok. If 750, it seems like a ridiculous handicap - aka an easy fight where I wouldn't take any damage controlling the fighter myself could suddenly becomes an auto-loss for the auto resolve. And isn't the point of auto-resolve to be able to skip those "easy" fights?

@Gauddlike - If there's multiple craft, if any of them are not able to roll then the game assumes none of them are able to roll. The damage output of the weapons doesn't need its own calculation, it should be reflected in the scores of the weapons in the existing system.

Canyou give us an example of how the system would resolve a battle between a Bomber and 2 Fighter's, vs 2 Condor's with Alenium missiles and Laser, and a Foxtrot with Alenium Torpedoes?

The damage output of the weapons doesn't need its own calculation, it should be reflected in the scores of the weapons in the existing system.

I was specifically talking about situations where a craft that couldn't win a fight in the players hands because it simply cannot do enough damage to kill the enemy can still win an autoresolve fight, even if it sometimes relies on a lucky rng.

For example a Corvette with 1700 health can be killed in autoresolve by a MiG armed with 2x alenium torpedoes despite those weapons having a maximum damage output of 1200 damage.

If the player was in control that would be an impossible fight to win and is only possible in auto resolve with a lucky rng roll.

A Marauder with a plasma cannon and 2x plasma torpedoes has a maximum damage of 2920 but an auto resolve strength of 3300 so can easily defeat a cruiser despite not having enough damage available to kill it if player controlled, this doesn't rely on lucky rng.

I am not sure if just tweaking numbers could resolve those problems as adding the airframe strength on top of the base damage of the weapons will almost always give a higher auto resolve value than the potential damage output of the whole craft.

That will always leave the door open to impossible wins.

I just don't know if that irritates other people as much as it does me

This isn't high enough to give you victory on its own so something is going wrong, but it's pretty close. I wonder if the UFOs are suffering from the fuel modifier or something themselves?

It's this.

I just did a test sending a Foxtrot with Alenium Torpedoes out against a Scout and the a Corvette. The autoresolver gave 85% chance to win against the Scout and 7% chance to win against the Corvette, which appear to be quite plausible values looking at the autoresolve values in the files.

Rather than fight, though, I withdrew the plane and sent it out again after a short while. Doing this a few times, the autoresolve chance eventually increased up to 100%.

(You could probably just fix this by increasing the range of UFOs from their present value of 10,000 up to 100,000 or 1,000,000 or something. Might be easier than coding special exceptions for UFOs?)

It's this.

I just did a test sending a Foxtrot with Alenium Torpedoes out against a Scout and the a Corvette. The autoresolver gave 85% chance to win against the Scout and 7% chance to win against the Corvette, which appear to be quite plausible values looking at the autoresolve values in the files.

Rather than fight, though, I withdrew the plane and sent it out again after a short while. Doing this a few times, the autoresolve chance eventually increased up to 100%.

(You could probably just fix this by increasing the range of UFOs from their present value of 10,000 up to 100,000 or 1,000,000 or something. Might be easier than coding special exceptions for UFOs?)

I'm not sure. The value remains constant for a while before depleting, like you'd expect from the player aircraft (a combat penalty is applied only when an aircraft has minimal fuel left). I think the issue is that the 'Range' variable for the UFOs doesn't actually affect their time in the air, but is still ticking down for the purposes of air combat. Therefore, when their fuel reaches '0' they can still fly around, but the autoresolve thinks it has no fuel and nerfs its combat value down to nothing.

Might be wrong, though. This is just the impression I got from the tests.

I'm not sure. The value remains constant for a while before depleting, like you'd expect from the player aircraft (a combat penalty is applied only when an aircraft has minimal fuel left). I think the issue is that the 'Range' variable for the UFOs doesn't actually affect their time in the air, but is still ticking down for the purposes of air combat. Therefore, when their fuel reaches '0' they can still fly around, but the autoresolve thinks it has no fuel and nerfs its combat value down to nothing.

Might be wrong, though. This is just the impression I got from the tests.

That sounds like a very possible explanation.

In my experience, engaging a UFO that has just entered the atmosphere gives a pretty good autoresolve value, but let it fly around for awhile and it you get 100% against it no matter what you bring to the fight.

That sounds like a very possible explanation.

In my experience, engaging a UFO that has just entered the atmosphere gives a pretty good autoresolve value, but let it fly around for awhile and it you get 100% against it no matter what you bring to the fight.

Use an iterative process. Each step use some ammo and fuel, apply damage to enemy and remove destroyed craft. Then repeat until all aircrafts or ufo are destroyed OR all ammo have been expended.

Actually, the easiest way to solve the problem of auto resolve killing enemies that you can't actually kill is to add an if statement that compares primary enemy HP vs damage potential.

Since escort's go away if the primary craft is killed, they shouldn't factor into that equation, but just add to the overall result score. May need some extra code so you don't get credit for killing them, however.

This way, the chassis score becomes a factor of aircraft maneuverability, helping to avoid taking damage in the fight.

I've noticed that, presumably since a weapon's ammo level only affects that weapon's strength and not the whole interceptor's, you can use the autoresolve to shoot down UFOs without actually carrying any ammo. I've been able to destroy a whole bunch of simultanous scouts with one Foxtrot that way, by scrambling it without waiting for it to rearm. That's another thing that would be fixed by Gauddlike's suggestion.

That shouldn't be possible, in more ways than one.

The Foxtrot only has a value of 300 without weapons while a scout has 1000.

Maybe the fuel level bug in action or another one showing up.

Plus the obvious one, what are they using to damage the enemy?

Unless all of the pilots went to the same flight school as Will Smith in Independence Day.

Unless all of the pilots went to the same flight school as Will Smith in Independence Day.

Everyone knows all xenonaut pilots are Will Smith clones. That is why you never have to hire any pilots. It is also the cause of the 100% autoresolve bug, the computer knows he always wins.

I've noticed that, presumably since a weapon's ammo level only affects that weapon's strength and not the whole interceptor's, you can use the autoresolve to shoot down UFOs without actually carrying any ammo. I've been able to destroy a whole bunch of simultanous scouts with one Foxtrot that way, by scrambling it without waiting for it to rearm. That's another thing that would be fixed by Gauddlike's suggestion.

What version are you playing?

I'm not sure that this is a problem, but it is interesting. At no point is a Corsair ever equal to a Foxtrot against an escorted UFO in autoresolve. The Corsair chassis is worth 300 more than a Fox, but the weapons considerably less. Laser vs Alenium is 1200 vs 1500. Plasma tech is 1600 vs 2100, and Mag/Fusion is 2100 vs 2700. Marauders outclass Corsairs by even more, add another 700 to the Fox score, then add another 300/500/750 depending on cannon tech (laser, plasma, mag).

Marauder vs Corsair, autoresolve version versus pure rolling craft then. I'll assume plasma tech since I generally have it before Marauders. 1600 (Corsair) vs 2000 (Marauder). 2100 vs 2500 at endgame. Condors actually beat Heavy Fighters 1 on 1 from laser/alenium on (800 to 700). Condors with plasma tech edge out Interceptors (1200 to 1150), but that's close enough to not be a sure win. One more weapon upgrade, and it goes hard in the Condor's favor (especially in 3 vs 3 scenarios.)

My conclusion would be that aside from autoresolving pure Interceptor flights if they appear before you have Marauders or fusion explosives, Corsairs are never the optimal tool per plane. You could possibly make an arguement for two Corsairs being more space effective than 3 Condors, but I haven't ran the math on it.

At no point is a Corsair ever equal to a Foxtrot against an escorted UFO in autoresolve.

Are you sure on that?

If any of the enemy ships can roll then all of the Foxtrots heavy missiles drop to zero autoresolve strength (ARS) score.

Even if you swap them for light missiles their ARS is halved if any enemy can roll.

That should mean that any time you have a single enemy craft that can roll the Corsair keeps its ARS at a minimum of 1200, assuming 2x laser cannon, while the Foxtrot drops to 300 using 2x alenium torpedoes or 600 if using 2x alenium missiles.

This disparity gets more pronounced as you get to higher tech levels.

Using MAG tier weapons a Corsair would always have 2100 ARS (2x MAG storm cannon) while a Foxtrot would still have 300 ARS with 2x fusion torpedoes or 1050 with 2x hyper missiles.

The Marauder would have 1750 with MAG storm cannon and 2x fusion torpedoes or 2500 with MAG storm and 2x hyper missiles for comparison.

*edit*

Added a couple of images for comparison purposes.

The first has the ARS of all aircraft at each tier, even a tier you wouldn't normally have them at.

There is a separate value for each tier against a craft that cannot roll and against one that can ®.

For interceptors with missiles there is also a version armed with heavy missiles (H) and a light missile version (L).

The second is the same but with the cost of the aircraft removed so it is purely the ARS of the weapon loadout.